Influenza imposes a substantial disease burden among the US elderly aged 65 or older, resulting in 33,000 deaths annually. To help reduce this serious disease burden, Healthy People 2020 sets the goal of a 90 percent influenza vaccination rate among the elderly, whose vaccination rate seems to have reached a plateau of ~ 70 percent during the early 2000s. Healthy People 2020 also aims to eliminate racial/ethnic disparities in vaccination rates among subgroups. This is because lower influenza vaccination rates are consistently reported among Hispanic and African American elderly, compared to non-Hispanic White elderly. Our findings will help policy-makers and healthcare leaders (I) allocate vaccine more efficiently in the short-term and (II) stabilize vaccine supply in the long-term. This will contribute to the Healthy People 2020 goals described above. Specifically, this study will predict influenza vaccine demand and identify key factors potentially contributing to racial/ethnic disparities in influenza vaccination. These factors include vaccine supply shortage and an individual behavioral response to avoid influenza-related risk (called avoidance response in this application), which can change during a single season. We will focus on vaccine supply shortage because our prior study revealed a strong dose-response association between more serious vaccine supply delay/shortage and greater racial/ethnic disparities among the US elderly population. Avoidance response is exemplified by an increase in influenza vaccine receipt after an observed increase in influenza epidemic activity. This time interval was empirically measured for both short-term (e.g., two weeks) and long-term (e.g., one year) time periods. Measuring avoidance response is useful in predicting both short-term and long-term influenza vaccine demand which may fluctuate, even within a single season, due to various causes. For instance, our model could predict state-specific influenza vaccine demand in the coming weeks based on the state-specific epidemic activity of the past weeks; this state-specific prediction could help improve short-term vaccine allocation across states. Our model will also predict long-term vaccine demand so healthcare providers and manufacturers will be better able to anticipate and meet the coming season's (long-term) demand. Further detailed predictions for each race/ethnicity group can be made with our model. The specific aims of the research plan are: 1) To assess whether racial/ethnic disparities in seasonal influenza vaccination worsened during the 2009-2010 season, in which serious seasonal influenza vaccine supply shortages occurred due to the pandemic (H1N1) influenza, compared to either the preceding (2008-2009) season or the subsequent (2010-2011) season. We will analyze the same cohort of Medicare beneficiaries aged e65 (enrolled in the nationally representative Medicare Current Beneficiary Survey (MCBS)). 2) To examine if the magnitude of avoidance response differs across racial/ethnic groups during a single season. 3) To develop a model for predicting the short and long-term demand for influenza vaccine, using results from aims 1 and 2.